distances in organizations: innovation in an r&d lab · 2019. 7. 22. · attracted towards those in...

British Journal of Management, Vol. 00, 1–16 (2015)DOI: 10.1111/1467-8551.12114

Distances in Organizations: Innovationin an R&D Lab

Wilfred Dolfsma and Rene van der Eijk1University of Groningen, School of Economics and Business, PO Box 800, 9700 AV Groningen, The

Netherlands, and 1RSM Erasmus University, Rotterdam, The NetherlandsCorresponding author email: [email protected]

The distance between actors in an organization affects how they interact with each other,and particularly whether they will exchange (innovative) knowledge with each other. Ac-tors in each other’s proximity have fewer conflicts, more trust towards each other, forexample, and are thus more involved in knowledge transfer. Actors close to others thusare believed to perform better: by being more innovative, for instance. This theory ofpropinquity’s claim resonates widely in the literature and has intuitive appeal: ‘people aremost likely to be attracted towards those in closest contact with them’ (Newcomb, Th.(1956). American Psychologist, 11, p. 575). Knowledge that a focal actor receives fromalters who are close is more readily accessed, better understood and more readily use-able. At the same time, however, and in contrast to the what the theory of propinquitysuggests, knowledge that a focal actor receives from alters who are at a greater distancemay be more diverse, offer unexpected and valuable insights, and therefore give rise toinnovation. In order to understand these opposing expectations, scholars have indicatedthat distance must be conceived of as multifaceted: individuals can be close to each otherin one way, while at the same time distant in another. No prior paper has extensivelystudied the effects of distance as a multifaceted concept, however. This study offers twodistinct contributions. It argues, first, why some instances of distance affect the oppor-tunity to interact with alters, potentially lowering an actor’s performance, while otherinstances of distance affect the expected benefits from interaction. The latter would in-crease an actor’s performance. Secondly, this paper is the first study to test empiricallythe expectations about how seven different measures of distance affect an actor’s innova-tive performance. Innovative performance is measured as both creative contribution andcontribution to knowledge that has immediate commercial use (patents). In the settingof a large research lab, it is found, contrary to expectations, that distance does not hurtindividual innovative performance and sometimes helps it in unexpected ways.

Distance between actors in an organization is be-lieved to affect whether they will interact with eachother to exchange knowledge (Akerlof, 1997). Inthe literature, interaction and knowledge exchangeare firmly expected to stimulate individual perfor-mance and innovativeness. The theory of propin-quity, as suggested by Newcomb (1956, p. 575),states clearly that ‘people are most likely to beattracted towards those in closest contact withthem’. In particular, the extent to which actors arelikely to exchange and build relations decreases asdistance between them increases (Akerlof, 1997).If knowledge is received from ‘distant’ others, it is

not likely to be readily accessed, understood andused (Dolfsma, Finch and McMaster, 2011). Be-cause of distance between individuals, there maynot be interaction or exchange of knowledge, andthe knowledge that is exchanged can bemore easilymisunderstood. Since innovation comes from thecombination of different pieces of knowledge, in-dividuals are thus less likely to be innovative if thedistance between them and others increases. Be-yond the effect of distance between individuals ontheir innovativeness,Monge et al. (1985) stress that‘a variety of organizational outcomes’ are affectedby distance between individuals.

© 2015 British Academy of Management. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX42DQ, UK and 350 Main Street, Malden, MA, 02148, USA.

2 W. Dolfsma and R. van der Eijk

This premise is a key one, particularly in a line ofresearch that focuses on the functioning of globalor virtual teams – a key topic in today’s globaliz-ing and competitive business environment (Cum-mings, 2004;Hinds andMortensen, 2005;Martins,Gilson and Maynard, 2004; Maznevski and Chu-doba, 2000; O’Leary and Cummings, 2007; Olsonand Olson, 2000). The idea in this line of researchis that ‘out of sight, means out of sync’ (Hinds andBailey, 2003).

Distance, however, is not a singular term, butcan have multiple dimensions, instantiations orfacets. Most ways in which distance has beenconceived and its consequences theorized, how-ever, assume that distance hampers knowledge ex-change and so negatively affects individual inno-vativeness and performance. Knowledge receivedfrom alters in one’s proximitymay be too similar tothe knowledge that one already has, while knowl-edge received from alters who are more distant ismore different and may lead to more actually newknowledge arising. Some suggest that the effect ofdistance on knowledge transfer and innovativenesscan be beneficial (Gilsing et al., 2008; Wuyts et al.,2005). When and why this would be so remains un-clear, however.

We make two key contributions in this paper.The first is conceptual. In addition to categoriz-ing different instantiations of distance, we arguewhy some instances of distance affect the oppor-tunity to interact with alters, potentially loweringan actor’s performance, while other instances ofdistance affect the expected benefits of interaction.The latter would increase an actor’s performance.Increased benefits expected from an individualexchanging knowledge with alters at a distancewould materialize as increases in individual inno-vativeness, while increasing distance between anindividual and their alters decreases the opportu-nities to interact and decreases innovativeness. Per-sonal affiliation distance among individuals maybe close, indicating that the opportunity for knowl-edge exchange is high. Spatial distance betweenindividuals may be large, lowering the opportu-nity for exchange (Alba and Kadushin, 1976). In-dividuals exchanging over greater distancesmay beable to access knowledge unavailable in their im-mediate environment, thus possibly providing in-sights that help their innovative performance. Thispaper, secondly, is the first study to test empiri-cally the expectations about how seven different in-stantiations of distance affect an actor’s innovative

performance. We find, contrary to expectations,that distance does not hurt individual innovativeperformance, and sometimes helps it in unexpectedways, as in the case of hierarchical distance. De-constructing the notion of ‘distance’, and recog-nizing that some kinds of distance mostly affectthe opportunity for exchange, while others mostlyaffect the expected benefits of exchange, allows usto show that (1) some forms of distance stimulateinnovation in an organization and other measuresdo not, (2) some measures of distance contributeto one type of proxy for innovation and not to an-other, and thus (3) how distance is conceptualizedand measured is not a mere methodological con-cern. We investigate these contentions for knowl-edge transfer between laboratory scientists, usingtheir innovative performance measure comprehen-sively as both creative contribution performanceand contribution to knowledge that has immedi-ate commercial use (patents).

Theory: distances in organizations

Despite being little conceptualized (cf. Lechner,1991; Wilson et al., 2008), distance betweenindividuals has been acknowledged to have ‘con-siderable influence on a variety of organizationaloutcomes’ (Monge et al., 1985). The impact is be-lieved to be mostly negative: distance decreasestrust between individuals, increases the likelihoodand effects of conflicts, and will make people inan organization interact less frequently (Hindsand Bailey, 2003; Hinds and Kiesler, 1995; Mongeet al., 1985). The performance of individuals dis-tanced from other individuals and of an organi-zation where individual employees are at a dis-tance to others suffers. In more recent years thefocus for this line of research has moved to thestudy of global or virtual teams, but the suggestedeffects remain (Cummings, 2004; DiStefano andMaznevski, 2000; Martins, Gilson and Maynard,2004; O’Leary and Cummings, 2007). In thesestudies, we submit, different instantiations or di-mensions of ‘distance’ are conflated, giving riseto results that are not readily interpretable froman academic or a managerial point of view. Al-though there is some acknowledgement that dif-ferent dimensions to distance may need to berecognized, each of which will affect communica-tion in general, and knowledge transfer in partic-ular (Boschma, 2005; Danson, 2000; Napier and

© 2015 British Academy of Management.

Distances in Organizations 3

Ferris, 1993), affecting a large number of orga-nizational performance outcomes (Monge et al.,1985), in empirical studies ‘distance’ has mostlybeen analysed for one dimension only: spatial dis-tance (Monge et al., 1985; Rogers and Larson,1984; Saxenian, 1991; Singh, 2005).

Some studies focus on cognitive distance (Gils-ing et al., 2008; Nooteboom, 2000), as it is clearthat even those who are co-located may not read-ily understand each other if individuals have, forinstance, different cognitive backgrounds. Studiesfocusing on cognitive distance suggest that cog-nitive distance can be beneficial: if two partieshave toomuch knowledge in common, they cannotlearn from each other. Some others have focusedon social distance (Agrawal, Kapur and McHale,2008), as communication using electronic meanshas grownmore common, and spatial distance canbe overcome using different technical means. Inline with this, even being in the same team or socialcommunitymay notmean that individuals actuallyinteract and exchange knowledge. Some, therefore,focused on network distance between individuals(Alba and Kadushin, 1976; Reagans andMcEvily,2003). In part, an absence of exchange betweenany two individuals may be due to a hierarchi-cal distance between them as well, as individualsmay not exchange with others beyond a faultlineprovided by differences in hierarchy, for instance(Bezrukova et al., 2009). Exchange of knowledgemay be reduced if potential exchange partners arein a supervisor–subordinate relation (cf. Aalbers,Dolfsma and Leenders, forthcoming).

Each of these instantiations of distance is moreor less established in the relevant literatures, eventhough the literatures are somewhat disconnectedso far. Few empirical studies, however, have in-cluded multiple measures for distance, with theexception of macro or inter-firm studies in thedomain of economic geography (Agrawal, Kapurand McHale, 2008; Breschi and Lissoni, 2009).Few studies have conceptualized why some formsof distance may be beneficial, and others may bedetrimental to knowledge exchange.

We submit that distance between exchangingparties can affect the opportunity for knowledgeexchange between distanced individuals, on theone hand, and the expected benefit from knowl-edge exchange between distanced individuals, onthe other hand. If there is no opportunity forknowledge exchange, none may occur; if there isno expected benefit of knowledge exchange, none

will be initiated. Acknowledging that distance mayhave multiple instantiations suggests that, whilecognitive distant can offer larger expected bene-fits, in other respect the distance between cogni-tively distant individuals may be large as well. Astudy that does not conceptually acknowledge andmethodologically include this possibility may at-tribute findings for the one distance measure in-cluded that are in actual fact caused by otherdistance measures. Reduced opportunities for ex-change may be compensated for by increased ex-pected benefits of knowledge exchange over a dis-tance. Not recognizing the different instantiationsof the concept of distance may leave these dynam-ics unnoticed.

Opportunity for knowledge exchange

Distance can, first of all, fail to provide an oppor-tunity for exchange.Actors may be separated by spatial distance

and, classically, this is shown to prevent themfrom interacting and exchanging (Boschma, 2005;Danson, 2000). In a classical study of communi-cation and transfer of knowledge in a laboratory,Allen (1977) finds that even relatively limited geo-graphical distance between actors can hamper ex-change. Individuals simply may not meet to learnabout each other’s projects and knowledge needs.Distance may have a relational dimension

(Amin and Cohendet, 2004; Danson, 2000;Boschma, 2005), and be felt by the focal actor orattributed to the relation of the focal actor withan alter (Wilson et al., 2008). Kogut and Zander(1992) point out that, with regard to the innova-tion development process and since the formationof new cooperative relationships is a laboriousprocess, existing social relationships are usuallyemployed in the innovation development process.Knowledge exchange is facilitated by a personalrelationship between people, as exchange of espe-cially tacit knowledge is believed to benefit fromintrinsic motivation, trust and relationship specificlearning effects (Ingram and Robert, 2000; Moranand Ghosal, 1996; Nahapiet and Ghosal, 1998;Osterloh and Frey, 2000; Powell et al., 1996, Star-poli, 1998; Tsai and Ghosal, 1998). Alternatively,then, a personal distance felt between individualsin an organization can prevent knowledge transferfrom occurring. Person-related distance can giverise to faultlines in an organization (Bezrukovaet al., 2009). A number of individual factors



relating to someone’s personality traits and per-sonal history have been suggested to affect whatmay be called the personal distance experiencedbetween actors communicating (Monge et al.,1985; Wilson et al., 2008). Age and gender areamong these (Bezrukova et al., 2009). Value ori-entations have also been mentioned as a factor todetermine personal distance between individuals.Larger personal distance between focal actorsand their alters will, ceteris paribus, negativelyaffect their exchange of knowledge and thus theirinnovative performance.

What Danson (2000) calls organizational dis-tance can also prevent exchange. Organizationaldistance can have two dimensions: (1) distancecreated by unit boundaries; and (2) distancesdue to hierarchy. Units boundaries in an or-ganization can create hurdles for knowledgeexchange, even when individuals are co-located(Gulati and Puranam, 2009). By creating organi-zational unit boundaries (distance), communica-tion within the unit is enhanced, but communi-cation between units, crossing unit boundaries, ismade more difficult. Knowledge transfer and com-munication across boundaries ‘can be character-ized by false starts, different interpretations anddisruptions’ (Reagans and McEvily, 2003, p. 247)as organizational boundaries can be actively main-tained or even policed (Llewellyn, 1994; Zuck-erman, 1999), just like boundaries for sciences(Gieryn, 1999), genres in art (DiMaggio, 1987,1997; Hsu, 2005), markets (Ruef and Patterson,2009) and ethnic groups (Barth, 1969). Identities,status and what knowledge is taken for granteddepend on boundaries (DiMaggio, 1997; Douglas,1966; Hsu and Hannan, 2005; White, 1992; Zuck-erman, 1999). The division of labour that resultsfrom establishing unit boundaries allows for spe-cialization, largely attributable to the enhanced ex-change of knowledge within each unit (Hansen,1999; Uzzi, 1997). Ties that cross unit boundariesare more difficult to establish or maintain (Aal-bers, Dolfsma and Leenders, forthcoming; Mac-donalds and Williams, 1993a,b). Knowledge thatcrosses unit boundaries, and the messenger thathas brought it, may actually be regarded with sus-picion (Dolfsma, Finch andMcMaster et al., 2011;Hsu, 2006). An individual who acts as a boundaryspanner or gatekeeper, as a conduit for knowledgeto transfer into an organizational unit, may thushelp the organization, yet be in a precarious posi-tion at the same time.

Another measure for organizational distancewould be the distance between individuals, possi-bly within the same unit, who differ in hierarchicalrank: organizational hierarchical distance (Napierand Ferris, 1993). Faultline theory (Bezrukovaet al., 2009) suggests that interactions and ex-change between individuals may be affected by thehierarchical distance, often perceived as a faultline,between them. Levels of trust are lower betweenindividuals from across faultlines creating this or-ganizational distance (Li and Hambrick, 2005;Postuma and Campion, 2009). Individuals aresaid to be more likely to communicate, exchangeknowledge and ultimately perform well in their or-ganization if no or little hierarchical distance thatconstitute a faultline exists between them (Bor-gatti and Cross, 2003; Jung, Chow and Wu, 2003;Napier and Ferris, 1993; Wilson et al., 2008). Evenwhen knowledge crosses a faultline, arguments orfacts are weighed differently if received from acrossa faultline (vanKnippenberg and Schippers, 2007),and the amount of knowledge moving between in-dividuals decreases.

People may not be co-located, may not be for-mally working in the same unit, or may not beof the same rank in the organization, and yetcommunicate with each other, as they have es-tablished network contacts with each other (Aal-bers, Dolfsma and Koppius, 2014; Amin and Co-hendet, 2004), reaching beyond what Reagans andMcEvily (2003, p. 247) call ‘institutional, organi-zational or social boundaries’, thus reducing one’sdistance to others with whom one may usefullycommunicate and exchange knowledge, which islikely to result in interaction with a ‘differentbody of knowledge’ (Reagans and McEvily, 2003,p. 247). In such communications, people can per-ceive proximity, yet be at a large distance in otherrespects, providing opportunities for knowledgetransfer. Wilson et al. (2008) refer to the possi-bility of two individuals being located far fromeach other, yet feeling close as the paradox of‘far-but-close’. This can lead to the exchange ofknowledge relevant for innovation (Wilson et al.,2008). With some, even if distant in other respects,a focal actor may be in direct contact and canexchange knowledge directly: direct network dis-tance is low when a focal actor is in immediateclose contact, with a diversity of others in an orga-nization, quick to access relevant knowledge fromdifferent sources. The knowledge acquired whenthis direct network distance is low will help the



focal actor to be more innovative (Aalbers, Dolf-sma and Koppius, 2014; Aalbers et al., 2013;Breschi and Lissoni, 2009; Burt, 2004; Hansen,1999; Sparrowe et al., 2001). Focal actors that arethus closely connected to many others, have betteropportunities to exchange, and will see their inno-vative performance enhanced (Borgatti and Cross,2003; Oh, Labianca and Chungh, 2006; Reagansand Zuckerman, 2001, 2003). Along similar linesof argumentation, focal actors may be able to tapinto knowledge in an organization, accessing whatis relevant for their innovative efforts, indirectly.By leveraging their direct contacts, focal actors canaccess knowledge possessed by third parties, at asomewhat larger network distance, which was ar-gued and found to benefit their innovative per-formance (Aalbers, Dolfsma and Leenders, forth-coming; Burt, 1992; Ingram and Roberts, 2000).An even more diverse knowledge base can then bedrawn on, from a larger subset of an organization’smembers, and one is thus able to have a better senseof what existing knowledge finds support withinthe organization, or what new knowledge a focalactor may offer would find such support. Also, fo-cal actors can cast a wider net, seeking to obtainknowledge to complement their own if they canaccess a larger number of alters indirectly, beingcloser to them. Even though actors are dependenton direct contacts to provide them with indirectknowledge that these may access, the focal actorcan try actively to obtain such knowledge.1

Distance affects the opportunities that exist foran individual to exchange knowledge with othersin the organization. We have distinguished six dif-ferent instantiations of distance that affect oppor-tunities for knowledge transfer:

1. spatial distance2. personal distance3. organizational unit boundary distance4. hierarchical distance5. network distance, direct6. network distance, indirect.

In the above, we have argued that, as distance be-tween a focal actor and alters increases in such a

1Burt (1992, 2004) focuses on the network as a whole,pointing to the favorable position of bridges connectingseparated groups. While these bridges can benefit fromtheir position, or even exploit it for their own benefit,in the argument Burt presents, such positions are givenrather than actively created by a focal actor.

way that the opportunities for knowledge exchangeare reduced in any of these six different ways, thefocal actor’s innovative performance is likely to de-crease. We thus propose:

P1: Increased distance from a focal actor to oth-ers that reduces the opportunities for knowledgeexchange decreases the actor’s innovative perfor-mance.

Expected benefit from knowledge exchange

Some have not claimed just that distance betweenindividuals hampers exchange, but have actuallydefined distance as that which hampers exchangebetween agents (Danson, 2000, p. 174). Accord-ingly, communication between actors in an orga-nizational setting may be impeded because of dif-ferences in the education enjoyed, and the skillsor experience accumulated (Borgatti and Cross,2003; Dougherty, 1992; Reagans and McEvily,2003). What is tacit knowledge for some, takenfor granted background knowledge that facilitatesthe exchange of innovative knowledge, may notbe equally tacit for others, perhaps making ex-change of knowledge more difficult (Hinds andMortensen, 2005).Others, however, expect and have found

favourable performance outcomes when collab-orating individuals cognitively are not in closeproximity. Cognitive distance between a focalactor and his or her alters can, indeed, make surethat what is exchanged actually is more likely tobe a valuable contribution to the knowledge thata focal actor already possesses, increasing thelikelihood that the focal actor is innovative. Awider variety of knowledge sources is drawn on(Aalbers, Dolfsma and Leenders, forthcoming;Burt, 1992; Ingram and Roberts, 2000; Reagansand McEvily, 2003; Woodman, Sawyer and Grif-fin, 1993), leading to a more judicial weighing ofwhat knowledge is used, even when the distantknowledge one has acquired is not actually used(Cramton andHinds, 2005;Williams andO’Reilly,1998), enhancing individual performance (Allen,1977). Exchanging knowledge with such alterswill help focal actors to understand and developtheir own knowledge is such a way that it alignsbetter with knowledge developed by others in theorganization. Focal actors who exchange withothers at a larger cognitive distance to them seethe use of the knowledge they themselves develop



in a larger context. Exchange with another at acognitive distance, in other words, helps actorsto become more innovative (Burt, 2004; Reagansand Zuckerman, 2001; Rodan and Galunic, 2004;Sparrowe et al., 2001). Knowledge exchanged withanother who is closer is more likely to be similarto that of the focal actor, adding less to what thefocal actor already knows (Gilsing et al., 2008;Wuyts et al., 2005).

P2: Increased distance from a focal actor to oth-ers that reduces the expected benefits of knowl-edge exchange increases the actor’s innovativeperformance.

Data and methodResearch site

The data were collected at a research and develop-ment (R&D) lab of a Dutch multinational chemi-cal company with offices and production facilitiesin 49 countries around the world (cf. Siggelkow,2007). This study is therefore a case study, withknown advantages and disadvantages associatedwith this type of research. Given the exploratorynature of studying the effects of multiple instan-tiations of organizational distance, this seems war-ranted. A number of distance variables for individ-ual employees from different organizations, even ifthey can be determined, do not make sense. Socialnetwork data for different organizations cannotbe aggregated meaningfully, for instance. Whilea cross-sectional empirical research design wouldin other circumstances increase representativeness,focusing here on a single organization is unavoid-able. Representativeness must be established by re-peating the study for other, preferably dissimilar,organizations to determine what effect organiza-tion or organizational field specific circumstanceshave.

The company, which has annual sales of over€8bn, operates across a broad spectrumof businessactivities, including nutritional and pharmaceuti-cal ingredients, performance materials and indus-trial chemicals. The company is structured into anumber of clusters, which are further subdividedinto fairly autonomous operating business groupsresponsible for product development, manufactur-ing and sales. In the recent past, the companyshifted away from offering bulk products towardsoffering specialty and higher value-added prod-

ucts. This shift resulted in an even stronger focuson technology and innovation, making research anintegral part of the company’s strategy. The com-pany commits a substantial percentage of its re-sources to R&D and undertakes numerous initia-tives to stimulate and improve innovativeness.

Management agreed to the use of a networkquestionnaire, tailored for the specific setting andadministered to a total of 195 lab researchers andlab managers. The target population representedall researchers (lab assistants, for example, were ex-cluded) and project managers employed by the twoparticipating R&D labs. The decision to include allresearch and project managers in the study meantthat our survey would achieve a complete view ofthe network of individuals involved in knowledgedevelopment and diffusion. An electronic surveywas distributed to this population of R&D lab re-searchers or engineers. Within network analysis,one-site, socio-centric research approaches are thestandard, since this type of research design allowsfor the identification of a clear network boundary(e.g. Krackhardt, 1990).

The survey was distributed to the target pop-ulation through intra-company mail from the of-fice of the R&D managers. The decision to sendthe survey via internal organization mail ratherthan from a university address served two pur-poses: signalling the company’s support and avoid-ing possible technical problems. After three weeks,approximately 55% of the R&D network sur-veys were returned. We then sent out a per-sonalized reminder in case of non-response and,subsequently, personally approached remainingnon-respondents. Our study thus achieved a 97%survey response rate for the target population inthree rounds and one month of surveying – a highresponse rate required by social network analysis(Wasserman and Faust, 1994).

Measures

Data were gathered using a standard surveymethod incorporating a name generator question(dyadic level data), and questions to characterizeboth a relationship and an individual (e.g. Mars-den, 1990). In answering the name generator ques-tion (‘Over the past 6 months are there any workrelated contacts from whom you regularly sought(research related) information and advice to en-hance your effectiveness as a researcher?’ [Yourmost valued work contacts]), each respondent was



Figure 1. Frequent relations in research laboratory

asked to list his or her key contacts, offering 14spaces, with the possibility for respondents to addmore contacts. We did not require that a contactcorroborate a tie. Rather than use self-reportedcontact, to calculate the network variables (below)and draw the network figure, we use an in-degreeapproach. Using in-degree measures of how oftena focal actor is mentioned as a contact, is more reli-able (Sparrowe et al., 2001; Tsai, 2001; WassermanandFaust, 1994). To obtain a better understandingof what the relevant network in this organizationlooks like, Figure 1 offers a visual representation ofthe structure of the network of contacts in the re-search laboratories.2 The connected lab scientistsshared 1111 relationships. Six individuals turnedout to be isolates. The variables are described be-

2Figure 1 only includes the 798 frequent (daily andweekly) interactions; using Multi-Dimensional Scalingtechniques nodes that were ‘more similar’ – listing one an-other and sharing the same alters – are positioned closertogether.

low, and a correlation table is provided in theAppendix.

Dependent variable

As suggested by Rodan and Galunic (2004), indi-vidual innovation performance was measured bymeans of a performance item, which asked man-agers, drawing from company records, carefullyto rate the researcher’s creativity over the last 6[‘To what extent is this person particularly cre-ative: someone to come up with novel and usefulideas?’; use a 1–5 scale, from weak to outstand-ing]. The use of this Idea Performance measure toascertain innovativeness followed the notion thatmeasurement of innovativeness at the individuallevel, as pointed out in the literature, often requiressupervisor (or peer) assessment (Amabile, 1996;Moran, 2005). In line with previous research, theassessment asked managers to assess behavioursrather than attitudes, for a specific period (cf. Tsui,



1984). Interviews with senior managers in the or-ganization suggested that line management wouldbe most appropriate for ascertaining a researcher’sindividual innovation performance, given their di-rect involvement with and formal responsibility torate these researchers. As the table with descriptivestatistics in the Appendix shows, subjective inno-vation performance varied considerably across the195 person lab. This indicates that managers canand do differentiate between the innovative contri-bution that individual lab scientists make. The ex-tent to which the supervisor’s evaluation is subjectto social pressures or the inclination to avoid con-flict, for instance, can thus be perceived as limited.The judgment, taken from company records, is notmerely ‘subjective’.

More objective, perhaps, is Patent Performance.In order to complement our individual-level data,we sought an alternative way ofmeasuring individ-ual innovativeness. Patents are granted for knowl-edge that is thought to have industrial or commer-cial application (Griliches, 1990). The applicationneeds to be spelled out in some detail in the patentapplication. The number of patents per researcherwas used as an admittedly less than perfect proxyfor innovative output. This approach is consistentwith the existing practice to measure via patents,in an indirect way, both the technological compe-tence of a firm (Narin et al., 1987) and produc-tivity for individual researchers (Bertin and Wy-att 1988). The number of patents scientists havebeen granted can have a significant impact on theircareers (Dietz et al., 2000), yet patenting is mo-tivated quite differently in different scientific do-mains, with immediate financial incentives playinga minor role (Sauermann et al., 2010). Since thenumber of patents applied for is cumulative overtime, controlling for tenure is warranted.

Alternatively, using two performance outcomemeasures as dependent variables offers the oppor-tunity to determine how robust the finding foreach is. The more subjective innovation measureof idea performance is statistically unrelated to themore objective innovation measure of patent per-formance, as the correlation table in the Appendixshows.

Distance variables

At the very least, what can be indicated is that dis-tance lacks a uniform meaning and has been con-ceptualized or used to signify different things: ge-

ographical, cognitive, organizational (unit bound-aries, hierarchy), network and personal distance.Based on network data of who exchanges inno-vative knowledge with whom, we determine howdifferent forms of distance contribute to an indi-vidual’s innovativeness in subjective (evaluation bysupervisor) as well as in objective (patent applica-tions) terms.

The boundary of departments may create op-portunities for joint production within a depart-ment or unit, but may also make cooperationacross business unit boundaries more difficult, forinstance from a formal point of view. Membershipof a business unit is a measure for organizationaldistance separate from other measures. In a way,therefore, the business unit can be conceived of asa measure of organizational distance. At the sametime, however, this measure cannot be changed bythe, possibly joint, actions of communicating indi-viduals. For this reason, we decided to include thismeasure for distance, as a dummy variable, in allthe models that we estimate, rather than alternat-ing this measure for distance as we do for the othermeasures of distance to obtain regression results.In this way, the business unit variable is actuallya control variable. The laboratory studied has twobusiness units. Based on company records, respon-dents could each be traced to their respective busi-ness unit (0= business unit A; 1= business unit B).Since this variable is a dummy variable, and sinceits effect may interfere with that of other variablesfor distance too, we have included it in six modelspecifications in Table 1, as if it were a quasi con-trol variable.

Effects found for a lab scientist’s innovativenessmay be erroneously attributed to a variable such ascentrality or unit membership, if in actual fact geo-graphical distance between individuals may be theexplanation (Monge et al., 1985). Given how com-mon facilities for employees are provided, we mea-sure geographical distance as a co-location of des-ignated workspaces on the same floor in the samebuilding.

The hierarchical position of the respondentswas included for its potentially explanatory powerwith regard to performance. Centrality in a net-work such as the knowledge transfer network can,but need not, be related to ego’s formal positionin the organization’s hierarchy. Data for our hi-erarchy measure of organizational distance weredrawn from company personnel records. The datawere used as a basis for our measure of hierarchal



Table 1. Distance and individuals’ innovativeness

Model

1a 1b 2a 2b 3a 3b 4a 4bPatents Creativity Patents Creativity Patents Creativity Patents Creativity

ControlsTenure 0.126* −0.232*** 0.096 −0.214*** 0.112 −0.203*** 0.103 −0.22***Gender −0.18** −0.26*** −0.212*** −0.258*** −0.143* −0.19*** −0.187** −0.235***Dept. size −0.002 −0.197*** −0.026 −0.208*** 0.054 −0.15* −0.013 −0.197***IndependentsBusiness unit (BU) −0.048 −0.085 −0.062 −0.101 0.002 −0.052 −0.037 −0.085Spatial dist. −0.083 −0.095Formal dist.: scientist vs sr. scientist 0.229*** 0.24***Formal dist.: scientist vs management 0.196** 0.163**Personal dist. 0.029 −0.049Cognitive dist.Network dist. (range)

R2 0.063 0.116 0.074 0.118 0.114 0.162 0.057 0.108Adj. R2 0.042 0.097 0.047 0.093 0.085 0.135 0.031 0.084Overall F 2.991 4.417 2.777 4.675 3.914 5.846 2.22 4.425

Model

5a 5b 6a 6b 7a 7bPatents Creativity Patents Creativity Patents Creativity

ControlsTenure 0.119 −0.219*** 0.097* −0.194** 0.128* −0.231***Gender −0.177** −0.238*** −0.205*** −0.244*** −0.153* −0.232***Dept. size −0.032 −0.196*** −0.008 −0.211*** 0.057 −0.127*IndependentsBusiness unit (BU) −0.057 −0.078 −0.041 −0.1 −0.123 −0.164**Spatial dist.Formal dist.: scientist vs sr. scientistFormal dist.: scientist vs managementPersonal dist.Cognitive dist. 0.159** −0.023Network dist. (direct) −0.095 0.085Network dist. (indirect) 0.246*** 0.276***

R2 0.081 0.107 0.072 0.119 0.111 0.177Adj. R2 0.056 0.082 0.046 0.093 0.086 0.154Overall F 3.216 4.34 2.722 4.689 4.424 5.37

Two tailed; ***, **, *: significant at 1, 5 and 10% levels.

level [scientist, senior scientist and science man-ager]. These possible values were converted into adummy variable [0 = scientist, 1 = senior scientist,2 = manager].

In line with Marsden and Campbell (1984) andBurt (1992), respondents were asked to reflect onthe personal bond with each of their alters. Thepersonal distance variable measures how the focalactor perceives to be personally close to his alters.[‘How close is your working relationship with theperson in question?’ Scale 1–5; 1 = very strong, 2= strong, 3 = neutral, 4 = weak, 5 = very weak].Building on a measure developed by Rodan and

Galunic (2004), respondents were asked to assessthe extent to which the knowledge base of the re-ported alter was similar or dissimilar to their own[‘How similar or different is your knowledge fromyour contact’s knowledge?’ Scale 1–4; 1 = verysimilar, 2 = similar, 3 = different, 4 = very differ-ent.] The measure for cognitive distance taps intothe idea that innovation is facilitated by bringingtogether different, though not too different, knowl-edge bases (Burt, 2004; Nooteboom, 2000; Pelledet al., 1999). The measure was reverse coded (i.e.4 was recoded as 1, etc.) so that a value increasereflected increased knowledge similarity.



A distance variables has been calculatedfrom the network data collected along the linesexplained above (see Aalbers, Dolfsma and Kop-pius, 2014). A focal actor’s position in the networkbrings it close to others if the tie strength of theconnections of the focal actor to a diverse set ofother actors, across expertise areas, provides himor her with direct network distance (Burt, 1984;Marsden, 1987; Reagans and McEvily, 2003).In particular, we adopt Reagans and McEvily’s(2003, p. 255) network range indicator, whichcaptures the extent to which individuals maintainweak ties in a diverse network across multiples(Granovetter, 1973). We measure the diversity ofa focal actor’s network contacts by the numberof ties that cross department boundaries. Thetwo business units included in our study togetherhave 24 departments. Indirect network distance ismeasured as two-step reach, the number of altersa focal actor has indirect access to in a network,through direct contacts.

Controls

Within the 24 departments in which lab scientistscollaborate closely, scale effects in research mayemerge. Following Tortoriello (2006), departmentsize was included to control for networking andexchange opportunities, only because of the sizeof the working group of lab scientists. Scores forthe independent variables could be an artefact ofworking in a larger department. Information aboutthe gender of the respondent, as a demographic at-tribute with possible explanatory value, was gath-ered using the survey instrument (dummy variable:female = 1, male = 0). As Bezrukova et al. (2009)indicate, faultlines, such as gender, can affect inter-actions within a group and performance outcomesfor groups. Respondents were asked to report theirtenure in the organization (years), as a possible ex-planation for performance. One may expect differ-ences in the way in which newcomers interact andperform, compared with those who are already so-cialized into an organization, having establishedrelations over time (Gundry, 1993). We decidedto use duration of a person’s tenure rather thanage, since company-specific experience and con-tacts are relevant. In addition, since patent inno-vativeness may have a cumulative element, in thatit is firm-specific, tenure is the appropriate controlvariable. This does treat individuals who have hada career prior to joining this firm similarly to engi-

neers who may just have graduated, however. Ageof the respondent was nevertheless gathered usingthe survey instrument (in years). Including age as avariable in the regressions had no statistical effect,while tenure did have a statistically significant ef-fect (Table 1). More importantly, however, tenureis known to affect communication patterns (Ahujaand Galvin, 2003).

Estimation

The descriptive statistics, provided in the Ap-pendix, do not indicate statistical problems thatwould require the use of more complex and lessstraightforwardly interpretable statistical regres-sion methods than OLS. Multicollinearity, statis-tically, is not an issue – VIF values are well belowacceptable levels. Despite this, we have opted toanalyse the effects of distance on individual inno-vative performance separately for conceptual rea-sons. Since the different distances are sometimesat odds, sometimes complimentary and sometimesoverlapping, and since their effects have not beenstudied in a single study, including different mea-sures for distance in an organization into a singleregression would leave the results difficult to inter-pret (Agrawal, Kapur and McHale, 2008).

Results

We find difference instantiations for distance tohave different and unexpected effects on individualinnovativeness in the knowledge-intensive contextof a research laboratory. Effects can differ betweenperceived creativity and patent application. One ismore objective, perhaps, and focuses on outcomes.The other can be more subjective and focuses onthe process of innovation.

Proposition 1 suggests that, when distance fromthe focal actor to others increases such that op-portunities for knowledge exchange decrease, indi-vidual innovativeness decreases. We have analysedthe effects of six such distance-related opportuni-ties for knowledge exchange. Since business unitmembership is a fundamental variable that bothcaptures distance in some sense, but is also a givenfor employees, we have included this variable in allthe models we estimate. Among the control vari-ables, business unit membership turns out not tohave an effect on innovativeness (models 1a, 1b).Organizational distance created by business unit



boundaries seems either to be irrelevant, or is over-come by lab scientists creating opportunities forexchange by reducing distance in other respects.The last suggestion may have some value in it,given that the beta for business unit in model 7b,where indirect network distance is added, is neg-ative and significant (β = 0.0164. p


on the other. Distance between individuals is gen-erally believed to hamper knowledge transfer andthus individual innovativeness. We show, however,for the seven different instantiations of distance in-cluded in our study that the effects can be quiteunexpected. We find that instantiations of distancethat some explicitly believe to hamper individ-ual innovativeness – most pertinently geographi-cal and hierarchical distance – actually stimulateknowledge transfer and innovation. Rather thanreducing the opportunity to exchange knowledgeand hamper innovation, these increase such op-portunities. In the case of hierarchical distance,the Merton effect may be involved, whereby thoselower in rank will actively seek to exchange withthose higher in rank, at an exchange rate that canbe unfavourable to those lower in rank, in orderto be seen in more favourable light (Dolfsma, vander Eijk and Jolink, 2009; Merton, 1968). Thefavourable effect for knowledge exchange and in-novation ofmore spatial distancemay be explainedby the more diverse information sources that spa-tially distant individuals can draw on, while theirspatial distance is overcome by the use of meansof exchange that reduce the distance between par-ties in other ways. Granovetter (1973) suggests thisimplicitly. Allen (1977) finds, for instance, that ge-ographical distance between communication part-ners can be overcome if they are personally andcognitively close (cf. Crane, 1972; De Solla Priceand Beaver, 1969; Wilson et al., 2008). Being at alarge spatial distance from another team membergeographically may also not be problematic if oneis able to reach the other, using technical means,because of close personal or network distance, en-gaging in ‘action at a distance’ with alters (En-sign, 2009; Lave and Wenger, 1991; Wenger andSnyder, 2000). Individuals can (seek to) overcomecognitive distance, by reducing network distance.Interactions between different instantiations ofdistance in an organization is left for future re-search, however.

One instantiation of distance expected to havea favourable effect on knowledge exchange andindividual innovativeness – cognitive distance –only has that effect on the outcome of innova-tion (patents) and not on the process of innova-tion (creativity). This is contrary to the argumentsused to support Proposition 2. Perhaps the invest-ment required of an individual to interact with oth-ers who are cognitively distant is at the expenseof someone’s immediate innovative contribution,as ranked by his or her superior. More research

is needed here. In particular, being in close net-work contact with others indirectly is favourablefor knowledge transfer and innovation. It would beuseful to determine what knowledge actually is ex-changed in this indirect manner, to establish whyindirect rather than direct contacts matter. Per-haps knowledge acquired indirectly through one’snetwork contacts is more likely also to be fromdifferent departments or from individuals ofhigher seniority (cf. Aalbers, Dolfsma and Leen-ders, forthcoming). These interactions effects are,however, impossible to explore in this paper, as weexplain below.

This is an exploratory study, bringing togetherfor the first time a number of different instanti-ations of distance, theorizing and exploring em-pirically how they affect an individual’s perfor-mance in terms of innovative contribution. Thispaper clearly has some limitations. For one, theeffect of the different measurements of distanceone can imagine may differ by context and de-pendent variable studied. Findings in a settingthat is less knowledge-intensive than an R&D labcould present a different picture (cf. Allen, 1977;Breschi and Lissoni, 2009; Monge et al., 1985).Causal claims could be firmer if an organizationwere studied over a longer period of time, andpanel data were available. More use could also bemade of qualitative data in a subsequent study, tohelp suggest causal mechanisms. Some may won-der about the use of a relatively low number ofobservations. We do, however, meet the stringentcriteria on the necessary response rate for a socialnetwork study (Aalbers and Dolfsma, 2015;Wasserman and Faust, 1994), and far exceed thenumber of observations used in other studies (cf.Aalbers and Dolfma, 2015).

Owing to these data limitations, however, we re-frain in this paper from a more complicated anal-ysis that posits either non-linear effects for eachinstantiation of distance, or moderation effectswhereby different instantiations of distance inter-act. The former have been alluded to in the lit-erature (e.g. Wuyts et al., 2005), but not empiri-cally explored. We have been unable conclusivelyto explore empirically the effects of interactions be-tween different instantiations of distance. The find-ings for interactions between the one measure forexpected benefit of distance, on the one hand, andthe measures for opportunity for exchange, on theother (available on request from the authors), donot show a consistent picture. We attribute this todata limitations.



Appendix: Correlation table

Variable Mean Std. dev. n 1 2 3 4 5 6

1 Tenure 15.5794 11.37251 189 12 Gender 0.2256 0.41908 195 −0.293** 13 Dept. size 2.4213 0.61441 195 −0.131 0.086 14 Business unit (BU) 0.2564 0.43777 195 0.056 −0.064 −0.282** 15 Two step reach (in-degree) 40.9572 20.89781 187 0.081 −0.143 −0.333** 0.371** 16 Cognitive dist. 2.6074 0.51319 192 −0.11 −0.027 0.102 0.067 0.024 17 Network range 0.9246 0.05612 187 −0.063 −0.021 0.033 0.039 −0.041 −0.0438 Personal dist. 3.5015 0.4135 192 −0.096 0.073 0.072 −0.063 −0.183* 0.484**9 Physical dist. 0.542 0.30588 187 −0.094 −0.093 0.013 −0.08 0.034 −0.01210 Formal dist:. scientist vs sr. scientist 0.1538 0.36173 195 −0.021 −0.162* 0.104 −0.023 0.067 0.03411 Formal dist.: scientist vs management 0.3231 0.46886 195 0.066 −0.006 −0.410** −0.104 0.321** −0.149*12 Innovation perf.: patents 1.8519 2.00777 189 0.155* −0.214** −0.029 −0.019 0.213** 0.14213 Innovation perf.: creativity 3.4433 0.93818 194 −0.126 −0.217** −0.155* −0.022 0.269** −0.023

Variable 7 8 9 10 11 12 13

1 Tenure2 Gender3 Dept. size4 Business unit (BU)5 Two step reach (in-degree)6 Cognitive dist.7 Network range 18 Personal dist. −0.021 19 Physical dist. 0.288** −0.062 110 Formal dist.: scientist vs. sr. scientist −0.153* 0.074 −0.087 111 Formal dist.: scientist vs. management 0.072 −0.255** 0.02 −0.295** 112 Innovation perf.: patents −0.094 0.008 −0.073 0.200** 0.114 113 Innovation perf.: creativity 0.075 −0.056 −0.024 0.178* 0.13 0.125 1

Two tailed; ***, **, *: significant at 1, 5 and 10% levels.

References

Aalbers, R. andW.Dolfsma (2015).Networks of Innovation. Lon-don: Routledge.

Aalbers, H. L., W. A. Dolfsma and O. Koppius (2013). ‘Individ-ual connectedness in innovation networks: on the role of indi-vidual motivation’, Research Policy, 42, pp. 624–634.

Aalbers, R., W. Dolfsma and O. Koppius (2014). ‘Rich ties andinnovative knowledge transfer within a firm’, British Journalof Management, 25, pp. 822–848.

Aalbers, R., W. Dolfsma and R. Leenders (forthcoming). ‘Verti-cal and horizontal cross-ties: benefits of cross-hierarchy andcross-unit ties for innovative projects’, Journal of ProductInnovation Management, forthcoming.

Agrawal, A., D. Kapur and J. McHale (2008). ‘How do spa-tial and social proximity influence knowledge flows? Evidencefrom patent data’, Journal of Urban Economics, 64, pp. 258–269.

Ahuja, M. K. and J. E. Galvin (2003). ‘Socialization in virtualgroups’, Journal of Management, 29, pp. 139–280.

Akerlof, G. A. (1997). ‘Social distance and social decisions’,Econometrica, 65, pp. 1005–1027.

Alba, R. A. and Ch. Kadushin (1976). ‘The intersection of socialcircles’, Sociological Methods and Research, 5, pp. 77–102.

Allen, T. J. (1977).Managing the Flow of Technology: TechnologyTransfer and the Dissemination of Technological Informationwithin the R&D Organization. Cambridge, MA: MIT Press.

Amabile, T.M. (1996).Creativity in Context: Update to the SocialPsychology of Creativity. Boulder, CO: Westview Press.

Amin, A. and P. Cohendet (2004). Architectures of Knowledge:Firms, Capabilities, and Communities. Oxford: Oxford Univer-sity Press.

Barth, F. (1969). Ethnic Groups and Boundaries: The Social Or-ganization of Culture Difference. Boston, MA: Little Brown.

Bezrukova, K., K. A. Jehn, E. L. Zanutto and S. M. B. Thatcher(2009). ‘Do workgroup faultlines help or hurt?’, OrganizationScience, 20, pp. 35–50.

Borgatti S. P. and R. Cross (2003). ‘A relational view of infor-mation seeking and learning in social networks’,ManagementScience, 49, pp. 432–445.

Boschma, R. (2005). ‘Proximity and innovation: a critical com-ment’, Regional Studies, 39, pp. 61–74.

Breschi, S. and F. Lissoni (2009). ‘Mobility of skilled work-ers and co-invention networks: an anatomy of localizedknowledge flows’, Journal of Economic Geography, 9,pp. 439–468.

Burt, R. S. (1984). ‘Network items and general social survey’,Social Networks, 6, pp. 293–339.



Burt, R. S. (1992). Structural Holes: The Social Structure of Com-petition. Cambridge, MA: Harvard University Press.

Burt, R. S. (2004). ‘Structural holes and good ideas’, AmericanJournal of Sociology, 110, pp. 349–399.

Cramton, D. and Hinds, C. (2005). ‘Subgroup dynamics in inter-nationally distributed teams: ethnocentrism or cross-nationallearning?’ Research in Organizational Behavior, 26, pp. 231–263.

Crane, D. (1972). Invisible Colleges: Diffusion of Knowledge inScientific Communities. Chicago, IL: University of ChicagoPress

Cummings, J. N. (2004). ‘Work groups, structural diversity, andknowledge sharing in a global organization’Management Sci-ence 50, pp. 352–364.

Danson,M.W. (2000). ‘On the analytical dimension of proximitydynamics’, Regional Studies, 34, pp. 169–180.

DeSolla Price, D. and D. Beaver (1969). ‘Collabora-tion in an invisible college’, American Psychologist, 21,pp. 1011–1018.

Dietz, J. S., I. Chompalov, B. Bozeman, E. O. Lane and J. Park(2000). ‘Using the curriculum vita to study the career paths ofscientists and engineers: an exploratory assessment’, Sciento-metrics, 49, pp. 419–442.

DiMaggio, P. (1987). ‘Classification in art’, American Sociologi-cal Review, 52, pp. 440–455.

DiMaggio, P. (1997). ‘Culture and cognition’, Annual Review ofSociology, 23, pp. 263–287.

DiStefano, J. J. andM. L.Maznevski (2000). ‘Creating value withdiverse teams in global management’, Organizational Dynam-ics, 29, pp. 45–63.

Dolfsma, W., J. Finch and R. McMaster (2011). ‘Identify-ing institutional vulnerability: the importance of language,and system boundaries’, Journal of Economic Issues, 45,pp. 805–818.

Dolfsma, W., R. van der Eijk and A. Jolink (2009). ‘On a sourceof social capital: gift exchange’, Journal of Business Ethics, 89,pp. 315–329.

Dougherty,D. (1992). ‘Interpretive barriers to successful productinnovation in large firms’, Organization Science, 3, pp. 179–202.

Douglas, M. (1966). Purity and Danger. London: Routledge.Ensign, P. C. (2009). Knowledge Sharing among Scientists. New

York, NY: Palgrave Macmillan.Gieryn, T. F. (1999). Cultural Boundaries of Science: Credibility

on the Line. Chicago, IL: University of Chicago Press.Gilsing, V., B. Nooteboom, W. Vanhaverbeke, G. Duysters, and

A. van den Oord (2008). ‘Network embeddedness and the ex-ploration of novel technologies: technological distance, be-tweenness centrality and density’, Research Policy, 37, pp.1717–1731.

Granovetter, M. (1973). ‘The strength of weak ties’, AmericanJournal of Sociology, 78, pp. 1360–1380.

Griliches, Z. (1990). ‘Patent statistics as economic indicators’,Journal of Economic Literature, 28, pp. 1661–1707.

Gulati, R. and P. Puranam (2009). ‘Renewal through reorgani-zation: the value of inconsistencies between formal organiza-tion’, Organization Science, 20, pp. 422–440.

Gundry, L. K. (1993). ‘Fitting into technical organizations: thesocialization of newcomer engineers’, IEEE Transactions inEngineering Management, 40, pp. 335–345.

Hansen, M. T. (1999). ‘The search-transfer problem’, Adminis-trative Science Quarterly, 44, pp. 82–111.

Hinds, P. and S. Kiesler (1995). ‘Communication across bound-aries: work, structure, and use of communication technologiesin a large organization’, Organization Science, 6, pp. 373–393.

Hinds, P. J. and D. E. Bailey (2003). ‘Out of sight, out of sync:understanding conflict in distributed teams’,Organization Sci-ence, 14, pp. 615–632.

Hinds, P. J. and M. Mortensen (2005). ‘Understanding conflictin geographically distributed teams: the moderating effects ofshared identity, shared context, and spontaneous communica-tion’, Organization Science, 16, pp. 290–307.

Hsu, G. (2005). ‘Evaluative schemas and the attention of criticsin the US film industry’, Industrial and Corporate Change, 15,pp. 467–496.

Hsu, G. (2006). ‘Jacks of all trades and masters of none: audi-ences’ reactions to spanning genres in feature film production’,Administrative Science Quarterly, 51, pp. 420–450.

Hsu, G. and M. T. Hannan (2005). ‘Identities, genres, and orga-nizational forms’, Organization Science, 16, pp. 474–490.

Ingram, P. and P.W.Roberts (2000). ‘Friendship among competi-tors in the Sydney hotel industry’, American Journal of Sociol-ogy, 106, pp. 387–423.

Jung, D. I., C. Chow and A. Wu (2003). ‘The role of transfor-mational leadership in enhancing organizational innovation’,Leadership Quarterly, 14, pp. 525–544.

Kogut, B. and U. Zander (1992). ‘Knowledge of the firm, com-binative capabilities, and the replication of technology’, Man-agement of Technology, 3, pp. 383–397.

Krackhardt, D. (1990). ‘Assessing the political landscape: struc-ture, cognition and power in organizations’, AdministrativeScience Quarterly, 35, pp. 342–369.

Lave, J. and E. Wenger (1991). Situated Learning: LegitimatePeripheral Participation. Cambridge: Cambridge UniversityPress.

Lechner, F. J. (1991). ‘Simmel on social space’, Theory, Cultureand Society, 8, pp. 195–201.

Li, J. and D. C. Hambrick (2005). ‘Factional groups: a new van-tage on demographic faultlines, conflict and disintegration inwork teams’, Academy of Management Journal, 48, pp. 794–813.

Llewellyn, S. (1994). ‘Managing the boundary: how accountingis implicated inmaintaining the organization’,Accounting, Au-diting and Accountability Journal, 7, pp. 4–23.

Macdonalds, S. and C. Williams (1993a) ‘Beyond the boundary:an information perspective on the role of the gatekeeper in theorganization’, Journal of Product Innovation Management, 10,pp. 417–427.

Macdonalds, S. and C. Williams (1993b), ‘The survival of thegatekeeper’, Research Policy, 23, pp. 123–132.

Marsden, P. V. (1987). ‘Core discussion networks of Americans’,American Sociological Review, 52, pp. 122–31.

Marsden, P. V. (1990). ‘Network data and measurement’, AnnualReview of Sociology, 16, pp. 435–463.

Marsden, P. V. and K. E. Campbell (1984). ‘Measuring tiestrength’, Social Forces, 63, pp. 377–401.

Martins, L. L., L. L. Gilson and M. T. Maynard (2004). ‘Vir-tual teams: what do we know and where do we go from here?’,Journal of Management, 30, pp. 805–835.

Maznevski, M. L. and K. M. Chudoba (2000). ‘Bridging spaceover time: global virtual team dynamics and effectiveness’,Or-ganization Science, 11, pp. 473–492.

Merton, R. K. (1968). ‘The Matthew effect in science’, Science,159, pp. 56–63.



Monge, P. R., L. W. Rothman, E. M. Eisenberg, K. I. Miller andK. K. Kirste (1985). ‘The dynamics of organizational proxim-ity’,Management Science, 31, pp. 1129–1141.

Moran, P. (2005). ‘Structural vs. relational embeddedness: socialcapital and managerial performance’, Strategic ManagementJournal, 26, pp. 1129–1151.

Moran, P. and S. Ghosal (1996). ‘Value creation by firms’,Academy of Management Proceedings, pp. 35–39.

Nahapiet, J. and S. Ghosal (1998). ‘Social capital, intellectualcapital, and the organizational advantage’, Academy of Man-agement Review, 23, pp. 242–266.

Napier, B. J. and G. R. Ferris (1993). ‘Distance in organizations’,Human Resource Management Review, 3, pp. 321–357.

Narin, F., Noma, E. and Perry, R. (1987). ‘Patents as indicatorsof corporate technological strength’, Research Policy, 16, pp.143–155.

Newcomb, Th. (1956). ‘The prediction of interpersonal attrac-tion’, American Psychologist, 11, pp. 575–586.

Nooteboom, B. (2000). ‘Learning by interaction: absorptive ca-pacity, cognitive distance and governance’, Journal ofManage-ment and Governance, 4, pp. 69–92.

Oh, H., G. Labianca and M.-H. Chungh (2006). ‘A multilevelmodel of group social capital’, Academy of Management Re-view, 31, pp. 569–582.

O’Leary, M. B. and J. N. Cummings (2007). ‘The spatial, tempo-ral, and configurational characteristics of geographic disper-sion in teams’,MIS Quarterly, 31, pp. 433–452.

Olson, G. M. and J. S. Olson (2000). ‘Distance matters’, HumanComputer Interaction, 15, pp. 139–179.

Osterloh, M. and B. Frey (2000). ‘Motivation, knowledge trans-fer, and organizational forms’, Organization Science, 11, pp.538–550.

Pelled, L. H., K.M. Eisenhardt andK. R. Xin (1999). ‘Exploringthe black box: an analysis of work group diversity, conflict andperformance’, Administrative Science Quarterly, 44, pp. 1–28.

Postuma, R. A. and M. A. Campion (2009). ‘Age stereotypes inthe workplace: common stereotypes, moderators, and futureresearch directions’, Journal of Management, 35, pp. 158–188.

Powell, W. W., K. K. Koput and L. Smith-Doerr (1996). ‘Inter-organizational collaboration and the locus of innovation: net-work of learning in biotechnology’, Administrative ScienceQuarterly, 41, pp. 116–145.

Reagans, R. and B. McEvily (2003). ‘Network structure andknowledge transfer: the effects of cohesion and range’,Admin-istrative Science Quarterly, 48, pp. 240–267.

Reagans, R. and E. Zuckerman (2001). ‘Networks, diversity andperformance: the social capital of R&D units’, OrganizationScience, 12, pp. 502–517.

Rodan, S. and C. Galunic (2004). ‘More than network struc-ture: how knowledge heterogeneity influences managerial per-formance and innovativeness’, Strategic Management Journal,25, pp. 541–562.

Rogers, E. and J. Larson (1984). Silicon Valley Fever. New York,NY: Basic Books.

Ruef, M. and K. Patterson (2009). ‘Credit and classifica-tion: the impact of industry boundaries in nineteenth-century America’, Administrative Science Quarterly, 54,pp. 486–520.

Sauermann, H., M. Roach and W. Zhang (2010). ‘Is there a“Taste for Science”? Individuals’ preferences and the organi-zation of science’, Working Paper, Georgia Institute of Tech-nology, Atlanta, GA.

Saxenian, A. (1991)., ‘The origins and dynamics of productionnetworks in the Silicon Valley’, Research Policy, 20, pp. 423–437.

Siggelkow, N. (2007). ‘Persuasion with case studies’, Academy ofManagement Journal, 50, pp. 20–24.

Singh, J. (2005). ‘Collaborative networks as determinants ofknowledge diffusion patterns’, Management Science, 51, pp.756–770.

Sparrowe, R. T., R. C. Liden, S. J. Wayne and M. L. Kraimer(2001). ‘Social networks and the performance of individualsand groups’, Academy of Management Journal, 44, pp. 316–325.

Starpoli, C. (1998). ‘Cooperation in R&D in the pharmaceuticalindustry’, Technovation, 18, pp. 13–23.

Tortoriello, M. and D. Krackhardt (2010). ‘Activating cross-boundary knowledge: the role of Simmelian ties in the gen-eration of innovations’, Academy of Management Journal, 53,pp. 167–181.

Tsai, W. (2001). ‘Knowledge transfer in intraorganizational net-works: effects of network position and absorptive capacity onbusiness unit innovation and performance’, Academy of Man-agement Journal, 44, pp. 996–1004.

Tsai,W. and S. Ghosal (1998). ‘Social capital and value creation’,Academy of Management Journal, 41, pp. 464–476.

Tsui, A. S. (1984). ‘A role-set analysis of managerial reputation’,Organizational Behavior and Human Performance, 34, pp. 64–96.

Uzzi, B. (1997). ‘Social structure and competition in interfirmnetworks’, Administrative Science Quarterly, 42, pp. 35–67.

Van Knippenberg, D. and M. C. Schippers (2007). ‘Work groupdiversity’, Annual Review of Psychology, 58, pp. 15–41.

Wasserman, S. and K. Faust (1994). Social Network Analysis:Methods and Applications. Cambridge: Cambridge UniversityPress.

Wenger, E. and W. Snyder (2000). ‘Communities of practice: theorganizational frontier’, Harvard Business Review, Jan.–Feb.,pp. 139–145.

White, H. (1992). Identity and Control. Princeton, NJ: PrincetonUP.

Williams, K. Y. and C. A. O’Reilly (1998). ‘Demography and di-versity in organizations’, Research in Organizational Behavior,20, pp. 77–140.

Wilson, J. M., M. B. O’Leary, A. Metiu and Q. R. Jett (2008).‘Perceived proximity in virtual work’,Organization Studies, 29,pp. 979–1002.

Woodman, R.W., J. E. Sawyer and R.W. Griffin (1993). ‘Towarda theory of organizational creativity’,Academy ofManagementReview, 18, pp. 293–321.

Wuyts, S., M. G. Colombo, S. Dutta and B. Nooteboom (2005).‘Empirical tests of optimal cognitive distance’, Journal of Eco-nomic Behaviour and Organization, 58, pp. 277–302.

Zuckerman, E. W. (1999). ‘The categorical imperative: securitiesanalysts and the legitimacy discount’, American Journal of So-ciology, 104, pp. 1398–1438.



Wilfred Dolfsma is a professor of innovation and strategy at the University of Groningen, School ofEconomics and Business. He studies cooperation between firms and people that allows for innovation.

Rene van der Eijk, PhD RSM Erasmus University, is a research associate at the University of Gronin-gen School of Economics and Business, as well as an entrepreneur. His research focuses on innovationand cooperation in networks.


distances in organizations: innovation in an r&d lab · 2019. 7. 22. · attracted towards those in...

Documents